Permalink Submitted by detuned on Wed, 07/10/2019 - 23:14.
The reason it keeps happening is because getting the score values is a complicated process. The testers send me emails with their times, I plug them into a complicated Google spreadsheet that then computes the scores. Then I export the spreadsheet to a TSV file and upload it to a Linux machine. I then run some code on the machine that looks at the correct fields in the TSV file (puzzle name, author name, scores) and outputs an RTF file that contains the text that should go in the booklet. I then download the RTF file to my laptop, and then I open Adobe InDesign (a page-layout program) and update the link to the new RTF file. Then I save the INDD (InDesign) file in case future updates are needed, and then export it to a PDF file, which then gets synchronized with the cloud so that Karel can get to it. He adds the password security and gets it onto the webserver.
If a tester sends me some testing times after the deadline I've asked for them, I have to make a decision whether to ignore their test results, or update the score values so that they're more accurate. If they do, I end up doing this whole process again. Sometimes if the number change is small I'll try to "cheat" by just hand-editing the RTF file. But that's more likely to lead to mistakes.
Permalink Submitted by Admin on Thu, 07/11/2019 - 07:28.
I take the answer keys together with regular expressions how to submit it and points distribution from Wei-Hwa's Excel sheet.
In case there is different distribution than in the final PDF it can happen, that there is discrepancy in the first release.
I try to check these but it can happen that I miss something.
Why it keeps happening
The reason it keeps happening is because getting the score values is a complicated process. The testers send me emails with their times, I plug them into a complicated Google spreadsheet that then computes the scores. Then I export the spreadsheet to a TSV file and upload it to a Linux machine. I then run some code on the machine that looks at the correct fields in the TSV file (puzzle name, author name, scores) and outputs an RTF file that contains the text that should go in the booklet. I then download the RTF file to my laptop, and then I open Adobe InDesign (a page-layout program) and update the link to the new RTF file. Then I save the INDD (InDesign) file in case future updates are needed, and then export it to a PDF file, which then gets synchronized with the cloud so that Karel can get to it. He adds the password security and gets it onto the webserver.
If a tester sends me some testing times after the deadline I've asked for them, I have to make a decision whether to ignore their test results, or update the score values so that they're more accurate. If they do, I end up doing this whole process again. Sometimes if the number change is small I'll try to "cheat" by just hand-editing the RTF file. But that's more likely to lead to mistakes.
GP7 points
I take the answer keys together with regular expressions how to submit it and points distribution from Wei-Hwa's Excel sheet.
In case there is different distribution than in the final PDF it can happen, that there is discrepancy in the first release.
I try to check these but it can happen that I miss something.
Best regards,
Karel (Admin)